Annual Survey of Secondary Distributors of Refined Petroleum Products (SRPP)

Detailed information for 2013

The purpose of this survey is to obtain information on the volume of refined petroleum products distributed by secondary distributors in Canada.

Data release - October 17, 2014

• Questionnaire(s) and reporting guide(s)
• Description
• Data sources and methodology
• Data accuracy

Description

The Annual Survey of Secondary Distributors of Refined Petroleum Products (ASSDRPP) is a cost recovery survey sponsored by Natural Resources Canada (NRCan) and Environment and Climate Change Canada (ECCC). It implements the recommendations from a report of findings stemming from a Policy Research Data Gap (PRDG) initiative (Economy Wide Enhancement of Energy Statistics) that investigated, among other things, ways to improve energy consumption statistics in the Report on Energy Supply and Demand (RESD). The RESD data are extensively used by ECCC to derive greenhouse gas inventory calculations and by NRCan's Office of Energy Efficiency to produce the Energy Efficiency Handbook. The RESD is also the primary source of information provided to the International Energy Agency (IEA) in order to meet Canada's several international reporting requirements, such as those under the Kyoto Protocol.

Secondary distributors of refined petroleum products (RPPs) are commercial entities that act as buyers and resellers between the supplier and end user for RPPs. Secondary distributors have emerged in the marketplace in the last few years due to the divestment, by the refineries, of the distribution aspects of their products. Refineries have instead been increasingly relying on fuel resellers/secondary distributors for the distribution of their products.

This structural change introduced in the distribution process of RPPs by secondary distributors is expected to have a significant impact on the energy demand statistics of RPPs published in the RESD. Statistics for energy consumption are collected from the suppliers of that energy, the refineries, with the Annual End Use of Refined Petroleum Products (AEND) survey, record number 2168. The refineries in the AEND questionnaire however, are unable to clearly allocate the energy demand by end-user as represented by secondary distributors. Instead, they allocate all sales made to secondary distributors to the Commercial and Institutional sector, a heterogeneous grouping of approximately 20 industries in the service sector.

This reporting practice used by the refineries has resulted in inflated energy consumption statistics in the Commercial and Institutional sector and understated energy consumption in other sectors of the economy. This problem of over-representation in energy consumption sectors is only going to become more pronounced over time, as refineries accelerate the divestment of the distribution of their products. This over-representation of energy consumption statistics in the RESD has far reaching policy implications.

Data from the ASSDRPP will be integrated with the data from the AEND.

Subjects

• Energy
• Petroleum products

Data sources and methodology

Target population

The target population consists of all motor gasoline, diesel fuel, heating fuel and heavy fuel resellers (either retail or wholesale) establishments operating in Canada for at least one day between January and December of a calendar year. Establishments mainly engaged in the retail or wholesale distribution and sales of crude oil, liquefied petroleum gases (LPG), aviation fuel, asphalt, and lubricating oils and greases are excluded from the target population of this survey. Also excluded from this survey are establishments covered by the AEND survey.

The survey population is the collection of all fuel dealers and wholesaling establishments from which the survey can realistically obtain information. The survey population will differ from the target population due to difficulties in identifying all the units that belong to the target population because of a possible lack of detailed information for some units, particularly small businesses with low sales levels.

The survey population is comprised of all statistical establishments coded to NAICS 454311, 454319 (excluding firewood dealers) and a portion of establishments coded to NAICS 412110 on Statistics Canada's Business Register.

• Additional documentation

Instrument design

The questionnaires used in the survey have been designed to minimize different interpretations. The survey forms were field tested with respondents to ensure the questions, concepts and terminology were appropriate. Statistics Canada's Questionnaire Design and Resource Centre (QDRC) performed qualitative tests of the questionnaire by conducting cognitive interviews with 15 small, medium and large size companies across the country.

Statistics Canada also consulted with the Canadian Oil Heat Association.

Sampling

This is a sample survey with a cross-sectional design.

The frame used for sampling is Statistics Canada's Business Register (BR). The statistical unit is the establishment. The survey population consists of approximately 600 establishments classified to NAICS 454311, 454319 (excluding firewood dealers) and a portion of establishments coded to NAICS 412110. We adopt the "ever-alive" notion in constructing the survey population, in the sense that business units that operated only part of the time during the reference period are also included.

To enhance the efficiency of data collection, the in-scope portion of the population is stratified by geography (province/territory) and by size (annual business revenue as contained in the BR and/or other data sources). Each geographic area is divided into a "take-all" stratum, consisting of all large units representing a large combined percentage (95%) of the total revenue, and a "take-some" stratum of the remaining small units. All business units in the "take-all" stratum are surveyed, whereas a random selection of the "take-some" units is sampled. To reduce response burden on small units, the smallest establishments in each geographic area are excluded from sampling.The final survey sample is enhanced by network sampling, in the sense that it includes all operating units in our population that belong to the same business enterprises that would respond for the surveyed units.

Data sources

Data collection for this reference period: 2013-12-16 to 2014-03-31

Responding to this survey is mandatory.

Data are collected directly from survey respondents.

The survey is conducted using the mail-out / mail-back questionnaire approach, as well as using Computer Assisted Telephone Interviews (CATI) for capture, edit and follow-up. Collection of the data is performed by Statistics Canada's Regional Offices.

The questionnaires are mailed to the respondents of the survey at the end of the calendar year. An automatic fax reminder is sent to non-reporters around 15 days after mailing out the questionnaires. A telephone contact is made with non-reporting companies 15 days after the first fax follow-up to discuss reporting delinquency and possible special arrangements. A second fax is sent to persistent non-reporters later on in the collection period before collection is closed.

Significant effort is spent trying to minimize non-response during collection. Methods used, among others, are interviewer techniques such as probing and persuasion, repeated re-scheduling and call-backs to obtain the information, and procedures dealing with how to handle non-compliant (refusal) respondents.

If data are unavailable at the time of collection, a respondent's best estimates are also accepted.

Respondents can report to the survey by fax or by mail. Information can be transmitted securely electronically or by telephone. In exceptional cases a company may not be able to comply with the legal reporting deadlines and special reporting arrangements are determined.

View the Questionnaire(s) and reporting guide(s) .

Error detection

The first step of the processing consists of applying a set of edits that has been developed to find out errors and inconsistencies for each questionnaire. The type of rules covered are mainly associated to equality or additivity constraints. Some verifications are also done for specific fuels to make sure that there are no invalid entries.

Two outlier detection methods are also implemented by using BANFF, the edit and imputation generalized system developed by Statistics Canada. Firstly, values of selected variables (totals only) are compared across records for the current collection period to pinpoint values that seem to be inconsistent. Secondly, when historical data are available, the observed variable's trend over time is studied and compared between the different records to identify the ones for which the values would have abnormally changed.

Imputation

When additivity constraints are not respected, a prorating adjustment is used to make the sum of the components equal the desired total.

If a complete non-respondent for the current reference period was a respondent the previous year, a historical imputation method is used. This method imputes the value of the same unit for the preceding survey cycle by correcting it according to the average variation of the variable for all responding records for which historical data are available. This method is referred to as BANFF.

Finally, when partial non-response is observed, a donor imputation method that establishes an average profile of the respondents in each stratum (geographic area and revenue size) is used to find replacement values for the missing data.

Estimation

Totals and coefficients of variation associated to the totals are calculated at the Canadian level for the variables of the survey with the help of the Generalized Estimation System of Statistics Canada. The calculation of the sampling error can be done directly since the sampling plan used is relatively simple.

Two adjustments were done to obtain the final weights used to calculate these totals. A simple adjustment to take into account non-response was first done at the stratum level. The weights of the units that were part of the take-some strata were also increased in order to take into consideration the small units excluded from sampling. This last adjustment was calculated according to the total revenue represented by these small units.

Quality evaluation

Some studies that compare the estimates of the survey with the ones from the previous cycle were done to evaluate their quality. The data are also confronted to the ones from the AEND that collect similar information.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects that could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

In order to prevent any data disclosure, confidentiality analysis is done using the Statistics Canada Generalized Disclosure Control System (G-Confid). G-Confid is used for primary suppression (direct disclosure) as well as for secondary suppression (residual disclosure). Direct disclosure occurs when the value in a tabulation cell is composed of or dominated by few enterprises while residual disclosure occurs when confidential information can be derived indirectly by piecing together information from different sources or data series.

Revisions and seasonal adjustment

This methodology does not apply to this survey.

Data accuracy

The unweighted non-response rate for this survey is greater than 70%. Efforts are made during data collection to encourage the large distributors to answer the survey in order to avoid a non-response bias. Moreover, imputation methods are carefully studied to make the imputed data as realistic as possible.

The systems used during processing are developed in advance and tested with data from the previous year or with preliminary data from the current year to reduce as much as possible the risk of processing errors.

Coefficients of variation are calculated with each calculated estimate to produce a quality indicator of these estimates.

The response rate for this survey is usually 70%.